EP1701251A1 - Reconfigurable circuit and configuration switching method - Google Patents
Reconfigurable circuit and configuration switching method Download PDFInfo
- Publication number
- EP1701251A1 EP1701251A1 EP05256002A EP05256002A EP1701251A1 EP 1701251 A1 EP1701251 A1 EP 1701251A1 EP 05256002 A EP05256002 A EP 05256002A EP 05256002 A EP05256002 A EP 05256002A EP 1701251 A1 EP1701251 A1 EP 1701251A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- operation unit
- flop
- flip
- unit
- operator
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 125
- 230000006870 function Effects 0.000 claims description 18
- 238000004891 communication Methods 0.000 claims description 13
- 238000012546 transfer Methods 0.000 claims description 13
- 230000003139 buffering effect Effects 0.000 description 4
- 238000013507 mapping Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03K—PULSE TECHNIQUE
- H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
- H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
- H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
- H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
- H03K19/17736—Structural details of routing resources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/76—Architectures of general purpose stored program computers
- G06F15/78—Architectures of general purpose stored program computers comprising a single central processing unit
- G06F15/7867—Architectures of general purpose stored program computers comprising a single central processing unit with reconfigurable architecture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/38—Concurrent instruction execution, e.g. pipeline or look ahead
- G06F9/3867—Concurrent instruction execution, e.g. pipeline or look ahead using instruction pipelines
- G06F9/3875—Pipelining a single stage, e.g. superpipelining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/38—Concurrent instruction execution, e.g. pipeline or look ahead
- G06F9/3885—Concurrent instruction execution, e.g. pipeline or look ahead using a plurality of independent parallel functional units
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/38—Concurrent instruction execution, e.g. pipeline or look ahead
- G06F9/3885—Concurrent instruction execution, e.g. pipeline or look ahead using a plurality of independent parallel functional units
- G06F9/3893—Concurrent instruction execution, e.g. pipeline or look ahead using a plurality of independent parallel functional units controlled in tandem, e.g. multiplier-accumulator
- G06F9/3895—Concurrent instruction execution, e.g. pipeline or look ahead using a plurality of independent parallel functional units controlled in tandem, e.g. multiplier-accumulator for complex operations, e.g. multidimensional or interleaved address generators, macros
- G06F9/3897—Concurrent instruction execution, e.g. pipeline or look ahead using a plurality of independent parallel functional units controlled in tandem, e.g. multiplier-accumulator for complex operations, e.g. multidimensional or interleaved address generators, macros with adaptable data path
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03K—PULSE TECHNIQUE
- H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
- H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
- H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
- H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
- H03K19/17724—Structural details of logic blocks
- H03K19/17728—Reconfigurable logic blocks, e.g. lookup tables
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03K—PULSE TECHNIQUE
- H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
- H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
- H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
- H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
- H03K19/1778—Structural details for adapting physical parameters
- H03K19/17792—Structural details for adapting physical parameters for operating speed
Definitions
- the present invention relates to a re-configurable circuit capable of realizing a variety of functions programmably, and more particularly relates to a programmable mutual connection configuration technique suitable for a data path in an operation unit.
- the re-configurable circuit has a structure in which a plurality of so-called operation units for processing data is provided in an array.
- Fig. 1 shows an example of the configuration of a re-configurable circuit.
- the re-configurable circuit comprises a plurality of clusters, which are connected by, for example, a crossbar switch or the like and enables data transfer between the clusters.
- One cluster comprises an ALU array unit (operator group 2).
- the ALU array unit comprises a plurality of operation units.
- the operation unit 10 usually comprises an ALU, a multiplier and the like.
- the cluster 1 comprises an operator group 2 (ALU array unit), configuration memory 3 (setting memory) and a sequencer 4.
- the operator group 2 comprises a data input unit 5, a data buffer unit 6, a data buffer control unit 7, an inter-operator network 8, data memory 9 and operation units 10.
- the data input unit 5 supplies externally inputted data to the data memory 9, each operation unit 10 and the like via the inter-operator network 8.
- the data input unit 5 comprises the data buffer unit 6.
- the data buffer unit 6 selects the buffering/non-buffering of externally inputted data by a control signal from the data buffer control unit 7.
- the data buffer control unit 7 receives configuration information from the configuration memory 3, transmits a control signal to the data buffer unit 6 as the control signal according to the configuration information, and selects buffering or non-buffering of the input data.
- the inter-operator network 8 is mutually connected with a variety of components (such as the data input unit 5, data memory 9, operation unit 10 and the like).
- the inter-operator network 8 enables the data transfer between a variety of components connected to the inter-operator network 8 according to configuration information generated based on externally supplied configuration data (data generated by compiling a program).
- the data memory 9 records data via the inter-operator network 8.
- the operation unit 10 is set so as to perform a function related to configuration information by the configuration information.
- the configuration memory 3 loads configuration data onto the configuration memory 3 from an external storage device for storing configuration data, which is not shown in Fig. 1, such as a PC or the like (for example, loads using the communication means of the PC).
- the configuration memory 3 comprises a configuration data loading unit, which is not shown in Fig. 1, and generates/outputs a configuration switching condition signal based on a condition establishing signal (such as a chip-select signal) mainly transmitted from the operation unit 10 of a variety of re-configurable components constituting the operator group 2.
- a condition establishing signal such as a chip-select signal
- the configuration switching condition signal is generated based on the condition establishing signal and configuration data from the configuration memory 3.
- the sequencer 4 generates the address of the configuration information to be subsequently read by the configuration data based on the switching condition signal.
- Fig. 2 shows the configurations of the configuration memory 3 and operation unit 10 of the re-configurable circuit. Next, the data processing of the operation unit 10 is described below.
- configuration information is transferred from the configuration memory 3 to each operation unit 10, and each operation unit 10 is set. At this moment, the configuration information also controls connection switching between operation units 10 to set the input data path of each operation unit 10.
- an analog radio frequency (RF) unit 103 down-converts a signal received from an antenna 101 in order to demodulate it, an analog baseband (BB) unit 103 A/D converts it and a digital BB unit 104 demodulates it.
- RF radio frequency
- BB baseband
- a digital BB unit 104 demodulates it.
- latency in order to realize, for example, an IEEE802.11a PHY exclusive circuit of the digital BB unit 104 by a re-configurable circuit, latency must be sufficiently low. Therefore, sometimes a wireless LAN process cannot be realized by a re-configurable circuit.
- Such a problem is caused by the data transfer speed (operation cycle) between operation units 10.
- the data transfer speed (operation cycle) between operation units 10.
- the data transfer speed from the first-stage operation unit 10 (operation unit 10 for receiving input data) to the most remote operation unit 10 (operation unit 10 for outputting data).
- the transfer speed decreases.
- the operation unit 10 includes a predetermined process delay (FF12: flip-flop) .
- FF12 flip-flop
- This process delay is always fixed regardless of the complexity (multiplication, addition, AND (logical product), OR (logical sum) and the like) of a command given to the operation unit 10. Therefore, even if the operation is the repetition of any simple process, the latency (the number of steps of FFs used until the process is completed after data is inputted) increases at every process of the operation unit 10.
- a re-configurable circuit on which three operation units 10 are mapped as shown in Fig. 4 is described as an example. Since a FF12 is permanently provided for the interface of the operation unit 10, process delay of three clocks always occurs. This has no relation to the contents of (type of) an operation process performed by the operation unit 10 and process delay increases as the number of operation units 10 increases.
- Patent reference 1 although input/output signals are transmitted between the operation unit 10 and a long-haul horizontal programmable mutual connection channel via a short-haul horizontal programmable mutual connection channel and a programmable switch, there is no special description of a mapping method for improving the process speed of the operation unit.
- Patent reference 1 Japanese Patent Application No. 2002-76883
- a command allocated to an operation unit can be processed within an operation cycle, the command is outputted to a connection destination operation unit (next operation unit) without going through a FF of the operation unit.
- the FF is only used if necessary, the latency is reduced.
- connection destination operation unit Furthermore, if the command of the connection destination operation unit is consecutively processed within the operation cycle, its latency can be minimized by skipping the FF of the connection destination operation unit.
- a re-configurable circuit for improving the data transfer speed (operation cycle) of the operation unit group can also be provided.
- One aspect of the present invention is an operation unit in a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information.
- the operation unit comprises an operator for applying an operation process to data inputted to the operation unit, a flip-flop for delaying the transfer of the input data to the operator and a selector for switching between a route for transferring the input data to the operator via the flip-flop and a route for transferring the input data to the operator skipping (i.e. without using) the flip-flop.
- the routes are switched according to the selector switching condition setting information.
- the selector can be switched by a crossbar switch.
- Another aspect of the present invention is a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information.
- the re-configurable circuit comprises an operator for applying an operation process to data inputted to the operation unit and a flip-flop for delaying the transfer of the input data to the operator.
- a selector for switching between a route for transferring the input data to the operator via the flip-flop and a route for transferring the input data to the operator, skipping the flip-flop, is provided outside the operation unit, and the selector switches the routes, according to the selector switching condition setting information contained in the configuration information.
- Another aspect of the present invention is an operation unit configuration switching method in a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information.
- a switching condition setting for switching between a route for a selector provided for the operation unit transferring data inputted to the operation unit to an operator provided for the operation unit via a flip-flop provided for the operation unit and a route the data to the operator, skipping the flip-flop is reflected in configuration information, and the routes are switched by controlling the selector.
- an operation process time corresponding to the operation contents of (type of operation performed by) the operation unit is calculated according to information about operation contents for setting the operation process contents of each of the plurality of operation units (i.e. configurations of operation units) and information about connection contents (i.e. connections between operation units) in order to perform an operation process, an operation unit to be connected first is selected based on the connection contents between the operation units, an accumulated process time (total or overall process time) is calculated based on the operation process times of the individual operation units in connection order starting from the first operation unit, and the accumulated process time is compared with the predetermined operation cycle. If the accumulated process time is below the operation cycle, a route is set so as not to use the flip-flop. If it exceeds the operation cycle, it is set so as to use the flip-flop, and also the accumulated process time is set to the operation process time of the currently selected operation unit.
- the operation process time can be calculated in relation to the number of data routes to (wires connected to) the operation unit.
- the operation process time can include time caused by temperature fluctuations.
- a command allocated to the operation unit is processed within the process time of an operation frequency (for example, within one clock)
- the command is outputted to the connection destination operation unit skipping the FF of the operation unit.
- the command of the connection destination operation unit is consecutively processed (can also be processed) within the (same) operation cycle (for example, within one clock)
- the command also skips the FF of the connection destination operation unit.
- the data transfer speed (operation frequency) of the operation unit group of the re-configurable circuit can be improved, and since there is no need to always use the FF of the operation unit, latency can be reduced.
- Fig. 5 shows the internal configuration of the operation unit 10 of the present invention.
- the operation unit 10 receives input data via a selector 11.
- the selector 11 switches, according to switching condition setting information contained in configuration information transferred from configuration memory (setting memory) 3 to select between passing the input data through an FF 12 and skipping the FF12.
- the input data is data externally inputted to the operation unit 10, such as the operation result of another operation unit, a signal externally inputted from an integrate circuit provided with a re-configurable circuit or the like.
- the input data is inputted to an operator 13 and outputted after being operated.
- the operator 13 operates using a combinatorial circuit or a sequential circuit. For example, the operator 13 performs operations, such as addition, multiplication, logical OR, logical AND and the like.
- Fig. 6 the operation units 10 each with the selector 11 are mapped in the configuration of Fig. 4.
- an operation unit 10 to be used is determined according to switching condition setting information contained in configuration information, the connection of each operation unit 10 is determined and the selector 11 is switched.
- the switching method of the selector 11 describe in the first preferred embodiment is described below.
- the contents (addition, multiplication, logical OR, logical AND or the like) of the operation process of the operation unit and connection between operation units (inter-operator network 8) are determined.
- the mapping method of the switching process contents of the selector 11 of the operation unit is described according to he flowchart shown in Fig. 8 (although actually there is wiring for connection according to configuration information between operation units, it is not shown in Fig. 9).
- step S41 N operation units 10 to be mapped are selected, and a number, such as 1 or the like is attached to an operation unit 10 to which data is inputted. The number is attached by a counter function or the like.
- step S42 an accumulated process time up to the i-th operation unit is calculated and is compared with a predetermined comparison process time.
- the summing of process times is described below with reference to Fig. 9.
- the operation process time of operation unit 1 is calculated based on an operation process time corresponding to a prepared operation process.
- step S44 accumulated process time: total value
- step S43 the selector 11 of an operation unit indicated by variable i is switched to a route not using the FF 12.
- step S44 the selector 11 of an operation unit indicated by variable I is switched to a route using the FF 12.
- step S45 it is checked whether all the switching of the selectors 11 of the operation units 1-N is completed. If all the selectors 11 of the N operation units are not switched, the process returns to step S41 and the processes are continued until all the selectors 11 of the N operation units are switched.
- information about selector switching can be reflected in configuration information by a program.
- the selector switching method described in the first preferred embodiment is described with reference to Fig. 10.
- the contents (addition, multiplication, logical OR, logical AND or the like) of the operation process of the operation unit and connection between operation units (inter-operator network) are determined.
- the operation process time correspondence table is a table (calculation expression) in which a process time is set for each one-bit width.
- a correspondence table (operation expression) in which addition, multiplication, logical OR and logical AND correspond to A, B, C and D nsec, respectively, is prepared.
- 10 nsec is set for the operation of one bit width
- 20 nsec is set for the operation of two bit width
- X nsec is set for the operation of N bit width.
- step S65 an operation process time correspondence table corresponding to a data bit width inputted to the i-th operation unit is referenced, and the i-th operation process time Ti is calculated.
- step S67 Tsum (total value) obtained by summing the accumulated process times and the comparison process time are compared.
- the comparison process time presets an operation cycle (Tclk).
- the operation cycle is a system clock cycle.
- a cycle other than the system clock cycle can also be used.
- step S68(yes) the process proceeds to step S68(yes). Otherwise, the process in step S69(no) is performed.
- step S68 the switching of the selector 11 of an operation unit indicated by variable i is set so as not to use the FF 12.
- step S69 the switching of the selector 11 of an operation unit indicated by variable i is set so as to use the FF 12.
- step S610 the current accumulated process time Tsum is discarded, and the operation process time Ti of the current operation unit is assigned to Tsum.
- step S611 it is checked whether the switching of all the selectors 11 of the operation units 1-N is completed. If the switch setting of all the selectors 11 of the N operation units is not completed, the process returns to step S64 and the processes are continued until the setting of all the selectors 11 of the N operation units is completed.
- selector switching information can be reflected in configuration information.
- the delay time of an operation process due to temperature can also be added besides time data for a bit width.
- selector switching can also be reflected in configuration data.
- connection of the selector 11 for switching the FF 12 of the operation unit can be modified by using a crossbar switch. Therefore, the selector 11 can also be provided outside the operation unit 10.
- the program used in the present invention can be executed by supplying it from memory, such as a ROM and a RAM, an external storage device and a portable storage device recording its program code to a computer (such as a personal computer, etc.) and making the computer read and execute the program code.
- the program code read from the storage medium can realize the new function of the present invention, and the portable storage medium on which the program code is recorded or the like also constitutes the present invention.
- the portable storage medium for providing the program code a floppy (registered trademark) disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a DVD-ROM, a DVD-RAM, a magnetic tape, a non-volatile memory card, a ROM card, variety of storage media on which the program code is recorded via a network connection device, such as electronic mail, personal computer communication or the like, (in other words, a communication line) can be used.
- the functions of the above-described preferred embodiments can also be realized by enabling an OS operating on the computer to execute a part of the actual process or the entire process, according to the instruction of the program code.
- the functions of the above-described preferred embodiments can also be realized by enabling a CPU or the like provided for a function extension board inserted in a computer or a function extension unit connected to a computer to execute a part of the actual process or the entire process, according to the instruction of a program code after the program code read from a portable storage medium is written onto memory provided for the function extension board or the function extension unit.
- the configuration information generated based on the program is stored in the appropriate area of configuration memory 3 by the communication means of the computer.
- Fig. 11 shows the basic configuration of the re-configurable circuit of the present invention, in which operation units are mapped.
- the operation unit represents an encircled area (group) in which data can be exchanged in high speed.
- it represents a range in which the process can be completed within one clock.
- the operation process in the group can be stably performed at high speed, by disposing an FF for group communication in the neighborhood of a group boundary as shown in Fig. 11.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Logic Circuits (AREA)
Abstract
Description
- The present invention relates to a re-configurable circuit capable of realizing a variety of functions programmably, and more particularly relates to a programmable mutual connection configuration technique suitable for a data path in an operation unit.
- Recently, a re-configurable circuit whose hardware can be re-configured by a program has been proposed. Generally the re-configurable circuit has a structure in which a plurality of so-called operation units for processing data is provided in an array.
- Fig. 1 shows an example of the configuration of a re-configurable circuit. The re-configurable circuit comprises a plurality of clusters, which are connected by, for example, a crossbar switch or the like and enables data transfer between the clusters. One cluster comprises an ALU array unit (operator group 2). The ALU array unit comprises a plurality of operation units. The
operation unit 10 usually comprises an ALU, a multiplier and the like. - The
cluster 1 comprises an operator group 2 (ALU array unit), configuration memory 3 (setting memory) and asequencer 4. - The
operator group 2 comprises adata input unit 5, a data buffer unit 6, a data buffer control unit 7, an inter-operator network 8,data memory 9 andoperation units 10. - The
data input unit 5 supplies externally inputted data to thedata memory 9, eachoperation unit 10 and the like via the inter-operator network 8. For example, thedata input unit 5 comprises the data buffer unit 6. In this case, the data buffer unit 6 selects the buffering/non-buffering of externally inputted data by a control signal from the data buffer control unit 7. The data buffer control unit 7 receives configuration information from theconfiguration memory 3, transmits a control signal to the data buffer unit 6 as the control signal according to the configuration information, and selects buffering or non-buffering of the input data. - The inter-operator network 8 is mutually connected with a variety of components (such as the
data input unit 5,data memory 9,operation unit 10 and the like). The inter-operator network 8 enables the data transfer between a variety of components connected to the inter-operator network 8 according to configuration information generated based on externally supplied configuration data (data generated by compiling a program). Thedata memory 9 records data via the inter-operator network 8. Theoperation unit 10 is set so as to perform a function related to configuration information by the configuration information. - The
configuration memory 3 loads configuration data onto theconfiguration memory 3 from an external storage device for storing configuration data, which is not shown in Fig. 1, such as a PC or the like (for example, loads using the communication means of the PC). Theconfiguration memory 3 comprises a configuration data loading unit, which is not shown in Fig. 1, and generates/outputs a configuration switching condition signal based on a condition establishing signal (such as a chip-select signal) mainly transmitted from theoperation unit 10 of a variety of re-configurable components constituting theoperator group 2. For example, the configuration switching condition signal is generated based on the condition establishing signal and configuration data from theconfiguration memory 3. Thesequencer 4 generates the address of the configuration information to be subsequently read by the configuration data based on the switching condition signal. - Fig. 2 shows the configurations of the
configuration memory 3 andoperation unit 10 of the re-configurable circuit. Next, the data processing of theoperation unit 10 is described below. - In order to set each
operation unit 10, configuration information is transferred from theconfiguration memory 3 to eachoperation unit 10, and eachoperation unit 10 is set. At this moment, the configuration information also controls connection switching betweenoperation units 10 to set the input data path of eachoperation unit 10. - According to
Patent reference 1, in a re-configurable device having a programmable mutual connecting network suitable for a data path, both input/output of signal transmission between a function cell (for setting a variety of logic functions programmably) and a long-haul horizontal programmable mutual connection channel are performed via a short-haul horizontal programmable mutual connection channel and a programmable switch. By such a configuration, the load of the long-haul horizontal programmable mutual connection channel can be reduced to realize high-speed transmission. A re-configurable device high-speed programmable mutual connecting network which secures sufficient routability using a little switching and wiring and especially in which a multi-bit data path can be efficiently implemented, is proposed. - However, in a system using a re-configurable circuit in which a plurality of
operation units 10 are provided in an array, for example, in the case of the wireless LAN receiving unit shown in Fig. 3, an analog radio frequency (RF)unit 103 down-converts a signal received from anantenna 101 in order to demodulate it, an analog baseband (BB) unit 103 A/D converts it and adigital BB unit 104 demodulates it. In this case, in order to realize, for example, an IEEE802.11a PHY exclusive circuit of thedigital BB unit 104 by a re-configurable circuit, latency must be sufficiently low. Therefore, sometimes a wireless LAN process cannot be realized by a re-configurable circuit. - Such a problem is caused by the data transfer speed (operation cycle) between
operation units 10. For example, in a structure where a plurality ofoperation units 10 are provided in an array, what matters is the data transfer speed from the first-stage operation unit 10 (operation unit 10 for receiving input data) to the most remote operation unit 10 (operation unit 10 for outputting data). In other words, as the number ofoperation units 10 for performing the operation process increases, its transfer speed decreases. - The
operation unit 10 includes a predetermined process delay (FF12: flip-flop) . This process delay is always fixed regardless of the complexity (multiplication, addition, AND (logical product), OR (logical sum) and the like) of a command given to theoperation unit 10. Therefore, even if the operation is the repetition of any simple process, the latency (the number of steps of FFs used until the process is completed after data is inputted) increases at every process of theoperation unit 10. - A re-configurable circuit on which three
operation units 10 are mapped as shown in Fig. 4 is described as an example. Since a FF12 is permanently provided for the interface of theoperation unit 10, process delay of three clocks always occurs. This has no relation to the contents of (type of) an operation process performed by theoperation unit 10 and process delay increases as the number ofoperation units 10 increases. - In
Patent reference 1, although input/output signals are transmitted between theoperation unit 10 and a long-haul horizontal programmable mutual connection channel via a short-haul horizontal programmable mutual connection channel and a programmable switch, there is no special description of a mapping method for improving the process speed of the operation unit. - Patent reference 1:
Japanese Patent Application No. 2002-76883 - In embodiments of the present invention, if a command allocated to an operation unit can be processed within an operation cycle, the command is outputted to a connection destination operation unit (next operation unit) without going through a FF of the operation unit. Thus, since the FF is only used if necessary, the latency is reduced.
- Furthermore, if the command of the connection destination operation unit is consecutively processed within the operation cycle, its latency can be minimized by skipping the FF of the connection destination operation unit.
- By grouping a plurality of operation units and using a FF for communication between groups when exchanging data between groups, a re-configurable circuit for improving the data transfer speed (operation cycle) of the operation unit group can also be provided.
- One aspect of the present invention is an operation unit in a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information. The operation unit comprises an operator for applying an operation process to data inputted to the operation unit, a flip-flop for delaying the transfer of the input data to the operator and a selector for switching between a route for transferring the input data to the operator via the flip-flop and a route for transferring the input data to the operator skipping (i.e. without using) the flip-flop. The routes are switched according to the selector switching condition setting information. The selector can be switched by a crossbar switch.
- Another aspect of the present invention is a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information. The re-configurable circuit comprises an operator for applying an operation process to data inputted to the operation unit and a flip-flop for delaying the transfer of the input data to the operator. A selector for switching between a route for transferring the input data to the operator via the flip-flop and a route for transferring the input data to the operator, skipping the flip-flop, is provided outside the operation unit, and the selector switches the routes, according to the selector switching condition setting information contained in the configuration information.
- In such configurations, by using a selector, a path without going through the FF can be selected, thereby reducing latency.
- Another aspect of the present invention is an operation unit configuration switching method in a re-configurable circuit provided with a plurality of operation units, capable of realizing a variety of functions by re-configuring the plurality of operation units, according to configuration information. In the method, a switching condition setting for switching between a route for a selector provided for the operation unit transferring data inputted to the operation unit to an operator provided for the operation unit via a flip-flop provided for the operation unit and a route the data to the operator, skipping the flip-flop, is reflected in configuration information, and the routes are switched by controlling the selector.
- In the selector switching condition setting, an operation process time corresponding to the operation contents of (type of operation performed by) the operation unit is calculated according to information about operation contents for setting the operation process contents of each of the plurality of operation units (i.e. configurations of operation units) and information about connection contents (i.e. connections between operation units) in order to perform an operation process, an operation unit to be connected first is selected based on the connection contents between the operation units, an accumulated process time (total or overall process time) is calculated based on the operation process times of the individual operation units in connection order starting from the first operation unit, and the accumulated process time is compared with the predetermined operation cycle. If the accumulated process time is below the operation cycle, a route is set so as not to use the flip-flop. If it exceeds the operation cycle, it is set so as to use the flip-flop, and also the accumulated process time is set to the operation process time of the currently selected operation unit.
- The operation process time can be calculated in relation to the number of data routes to (wires connected to) the operation unit. Preferably the operation process time can include time caused by temperature fluctuations.
- If a command allocated to the operation unit is processed within the process time of an operation frequency (for example, within one clock), the command is outputted to the connection destination operation unit skipping the FF of the operation unit. If the command of the connection destination operation unit is consecutively processed (can also be processed) within the (same) operation cycle (for example, within one clock), the command also skips the FF of the connection destination operation unit. By predetermining a process time for each operation process (command), the data transfer speed (operation frequency) of the operation unit group of the re-configurable circuit can be improved. In this case, there is no need to always use the FF of the operation unit, thereby reducing latency.
- Furthermore, if the sum of the operation process times of the operation units is below a predetermined operation cycle when the operation units each with a switching unit for performing the switching are connected according to the configuration information, the re-configurable circuit, being one aspect of the present invention comprises a flip-flop for communication at the output of the group and a communication unit for transferring data via the flip-flop for communication when transferring data between the groups.
- By such a configuration, by grouping several operation units and using a FF for communication between groups when exchanging data between groups, the data transfer speed (operation frequency) of the operation unit group can be improved.
- According to the present invention, the data transfer speed (operation frequency) of the operation unit group of the re-configurable circuit can be improved, and since there is no need to always use the FF of the operation unit, latency can be reduced.
- Reference is made, by way of example only, to the accompanying drawings in which:
- Fig. 1 shows the configuration of a cluster.
- Fig. 2 shows the main elements of an embodiment of the present invention in the configuration of a cluster.
- Fig. 3 shows a circuit in which a re-configurable circuit is applied to a wireless LAN device.
- Fig. 4 shows the conventional re-configurable circuit (process delay: three clocks).
- Fig. 5 shows the configuration of an operation unit embodying the present invention.
- Fig. 6 shows connections between operation units in an embodiment of the present invention.
- Fig. 7 shows a process not using an FF (process delay: two clocks).
- Fig. 8 is the flowchart of the selector switching method of a first preferred embodiment.
- Fig. 9 shows the basic configuration of the re-configurable circuit for explaining the flowchart of the selector switching method of the first preferred embodiment.
- Fig. 10 is the flowchart of the selector switching method of a second preferred embodiment.
- Fig. 11 shows the basic configuration of the re-configurable circuit in the case where data is exchanged between groups.
- The preferred embodiments of the present invention are described in detail below with reference to the drawings.
- Fig. 5 shows the internal configuration of the
operation unit 10 of the present invention. Theoperation unit 10 receives input data via aselector 11. Theselector 11 switches, according to switching condition setting information contained in configuration information transferred from configuration memory (setting memory) 3 to select between passing the input data through anFF 12 and skipping the FF12. In this case, the input data is data externally inputted to theoperation unit 10, such as the operation result of another operation unit, a signal externally inputted from an integrate circuit provided with a re-configurable circuit or the like. Then, the input data is inputted to anoperator 13 and outputted after being operated. In this case, theoperator 13 operates using a combinatorial circuit or a sequential circuit. For example, theoperator 13 performs operations, such as addition, multiplication, logical OR, logical AND and the like. - In Fig. 6, the
operation units 10 each with theselector 11 are mapped in the configuration of Fig. 4. In the configuration of Fig. 6, anoperation unit 10 to be used is determined according to switching condition setting information contained in configuration information, the connection of eachoperation unit 10 is determined and theselector 11 is switched. - In Fig. 7, if a process time between "or" (operation unit: A) and "+" (operation unit 10: B) is below an operation cycle, the
selector 11 at the second stage (operation unit 10: B) is switched to skip the FF. By such a configuration, the conventional three clocks can be reduced to two clocks. It is OK if theselector 11 can be switched according to switching condition setting information contained in configuration information. - The switching method of the
selector 11 describe in the first preferred embodiment is described below. In order to generate switching condition setting information contained in configuration information, the contents (addition, multiplication, logical OR, logical AND or the like) of the operation process of the operation unit and connection between operation units (inter-operator network 8) are determined. For example, assuming that each operation unit is mapped as shown in Fig. 9, the mapping method of the switching process contents of theselector 11 of the operation unit is described according to he flowchart shown in Fig. 8 (although actually there is wiring for connection according to configuration information between operation units, it is not shown in Fig. 9). - In step S41,
N operation units 10 to be mapped are selected, and a number, such as 1 or the like is attached to anoperation unit 10 to which data is inputted. The number is attached by a counter function or the like. When the selector of thefirst operation unit 10 is switched, i=1 is set and stored. When the selector of asubsequent operation unit 10 is switched, the subsequent operation unit is selected by incrementing the variable i of a counter to i= i+1 after completing steps up to a determination process S46, which is described later. - In step S42, an accumulated process time up to the i-th operation unit is calculated and is compared with a predetermined comparison process time. The summing of process times is described below with reference to Fig. 9. In a plurality of operation units (N), it is assumed that
operation unit 1 is an operation unit in the case of i=1. The operation process time ofoperation unit 1 is calculated based on an operation process time corresponding to a prepared operation process. - For example, a correspondence table (table, operation expression) in which addition, multiplication, logical OR and logical AND correspond to A, B, C and D nsec, respectively. If the command of
operation unit 1 is addition, A nsec is stored. Then, if in i=2, the command ofoperation unit 2 is multiplication, B nsec is stored, and an accumulated process time (A+B), being the sum of the operation process time A nsec ofoperation unit 1 and the operation process time B nsec ofoperation unit 2, is calculated, and the accumulated process time is assigned to a variable "sum". - Then, "sum" (accumulated process time: total value) obtained by summing operation process times is compared with the comparison process time. The comparison process time sets an operation cycle (Tclk). In this case, the operation cycle is a system clock cycle. However, a cycle other than the system clock cycle can also be used. Then, if the accumulated process time is below the comparison process time, the process proceeds to step S43 (yes). Otherwise, the process in step S44 (no) is performed.
- In step S43, the
selector 11 of an operation unit indicated by variable i is switched to a route not using theFF 12. - In step S44, the
selector 11 of an operation unit indicated by variable I is switched to a route using theFF 12. - In step S45, it is checked whether all the switching of the
selectors 11 of the operation units 1-N is completed. If all theselectors 11 of the N operation units are not switched, the process returns to step S41 and the processes are continued until all theselectors 11 of the N operation units are switched. - In Fig. 9, when all the operation process times of operation units 1-1 are summed and the accumulated process time is compared with the predetermined comparison process time (Tclk), the accumulated process time exceeds the comparison process time. Therefore, The
FF 12 ofoperation unit 1 is used. Then, a subsequent operation unit m is selected, and in step S41 i=m is set. In step S42, the operation process time of operation unit m is selected from the correspondence table and is assigned to "sum". Then, the comparison process time and the accumulated process time (sum=only the operation process time of operation unit m) are compared, and it is determined whether the FF is used. If the selector settings of all the operation units are completed, the process terminates. - As described above, information about selector switching can be reflected in configuration information by a program.
- The selector switching method described in the first preferred embodiment is described with reference to Fig. 10. In order to generate switching condition setting information contained configuration information, the contents (addition, multiplication, logical OR, logical AND or the like) of the operation process of the operation unit and connection between operation units (inter-operator network) are determined.
- In step S61, the number i allocated to the operation unit is initialized and i=0 is set. In this case, i corresponds operation units 0-N. It is assumed that the
FF 12 of the i=0-th operation unit is used. The accumulated process time is initialized and Tsum=0 is set. - In step S62, an operation process time correspondence table corresponding to a data bit width inputted to the i=0-th operation unit is referenced. In this case, the operation process time correspondence table is a table (calculation expression) in which a process time is set for each one-bit width. For example, a correspondence table (operation expression) in which addition, multiplication, logical OR and logical AND correspond to A, B, C and D nsec, respectively, is prepared. Furthermore, in the case of addition, 10 nsec is set for the operation of one bit width, 20 nsec is set for the operation of two bit width and X nsec is set for the operation of N bit width.
- In step S63, an accumulated process time between the output of the FF of the operation unit and the output of the operation unit (after operation) is calculated based on the operation process time correspondence table, and Tsum=Ti(i=0) is set.
- In step S64, i is incremented, and i=i+1 is set. Then, a subsequent operation unit is selected.
- In step S65, an operation process time correspondence table corresponding to a data bit width inputted to the i-th operation unit is referenced, and the i-th operation process time Ti is calculated.
- In step S66, the i-th operation process time of the currently selected operation unit is added to the accumulated process time, and Tsum=Tsum+Ti is set.
- In step S67, Tsum (total value) obtained by summing the accumulated process times and the comparison process time are compared. The comparison process time presets an operation cycle (Tclk). In this case, the operation cycle is a system clock cycle. However, a cycle other than the system clock cycle can also be used.
- Then, if the total value is below the comparison process time, the process proceeds to step S68(yes). Otherwise, the process in step S69(no) is performed.
- In step S68, the switching of the
selector 11 of an operation unit indicated by variable i is set so as not to use theFF 12. - In step S69, the switching of the
selector 11 of an operation unit indicated by variable i is set so as to use theFF 12. In step S610, the current accumulated process time Tsum is discarded, and the operation process time Ti of the current operation unit is assigned to Tsum. - In step S611, it is checked whether the switching of all the
selectors 11 of the operation units 1-N is completed. If the switch setting of all theselectors 11 of the N operation units is not completed, the process returns to step S64 and the processes are continued until the setting of all theselectors 11 of the N operation units is completed. - By the above-described program, selector switching information can be reflected in configuration information.
- When calculating the operation process time of an operation unit, the delay time of an operation process due to temperature can also be added besides time data for a bit width.
- If one operation unit is connected to a plurality of operation units, in other words if two or more routes exist, by using the above-described program, selector switching can also be reflected in configuration data.
- Furthermore, the connection of the
selector 11 for switching theFF 12 of the operation unit can be modified by using a crossbar switch. Therefore, theselector 11 can also be provided outside theoperation unit 10. - In this case, the program used in the present invention (program shown in the flowchart of Fig. 8 or 10) can be executed by supplying it from memory, such as a ROM and a RAM, an external storage device and a portable storage device recording its program code to a computer (such as a personal computer, etc.) and making the computer read and execute the program code.
- In this case, the program code read from the storage medium can realize the new function of the present invention, and the portable storage medium on which the program code is recorded or the like also constitutes the present invention. For the portable storage medium for providing the program code, a floppy (registered trademark) disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a DVD-ROM, a DVD-RAM, a magnetic tape, a non-volatile memory card, a ROM card, variety of storage media on which the program code is recorded via a network connection device, such as electronic mail, personal computer communication or the like, (in other words, a communication line) can be used.
- In addition to enabling a computer to execute a program code that the computer reads onto memory, the functions of the above-described preferred embodiments can also be realized by enabling an OS operating on the computer to execute a part of the actual process or the entire process, according to the instruction of the program code.
- Furthermore, the functions of the above-described preferred embodiments can also be realized by enabling a CPU or the like provided for a function extension board inserted in a computer or a function extension unit connected to a computer to execute a part of the actual process or the entire process, according to the instruction of a program code after the program code read from a portable storage medium is written onto memory provided for the function extension board or the function extension unit.
- The configuration information generated based on the program is stored in the appropriate area of
configuration memory 3 by the communication means of the computer. - Fig. 11 shows the basic configuration of the re-configurable circuit of the present invention, in which operation units are mapped. The operation unit represents an encircled area (group) in which data can be exchanged in high speed.
- For example, it represents a range in which the process can be completed within one clock.
- If such a group operates together with another group requiring a high-speed process (if data is exchanged between groups), the operation process in the group can be stably performed at high speed, by disposing an FF for group communication in the neighborhood of a group boundary as shown in Fig. 11.
- The application of the present invention is not limited the above-described preferred embodiments, and a variety of improvements and modifications are possible.
Claims (8)
- An operation unit (10) in a re-configurable circuit (1) provided with a plurality of operation units (10), capable of realizing a variety of functions by re-configuring the plurality of operation units (10) according to configuration information, comprising:an operator (13) for applying an operation process to data inputted to the operation unit (10);a flip-flop (12) for delaying the transfer of the data inputted to the operator (13); anda selector (11) for switching between a route for transferring the data inputted to the operator (13) via the flip-flop (12) and a route for transferring the data inputted to the operator (13) skipping the flip-flop (12),the selector (11) being responsive to switching condition setting information, contained in the configuration information, to switch the routes.
- The operation unit (10) according to claim 1, wherein
the selector (11) is switched by a crossbar switch. - A re-configurable circuit (1) provided with a plurality of operation units (10), capable of realizing a variety of functions by re-configuring the plurality of operation units (10) according to configuration information, comprising:an operator (13) for applying an operation process to data inputted to the operation unit (10); anda flip-flop (12) for delaying the transfer of the data inputted to the operator (13); anda selector (11) for switching between a route for transferring the data inputted to the operator (13) via the flip-flop (12) and a route for transferring the data inputted to the operator (13) skipping the flip-flop (12),the selector (11) being responsive to switching condition setting information, contained in said configuration information, to switch between said routes.
- An operation-unit configuration switching method of a re-configurable circuit (1) provided with a plurality of operation units (10), capable of realizing a variety of functions, in which
a switching condition that a selector (11) provided for each operation unit (10) switches between a route for transferring the input data to the operator (13) via a flip-flop (12) and a route for transferring the input data to the operator (13) skipping the flip-flop (12) is reflected in configuration information, and
the selector (11) is controlled to switch the routes, based on the switching condition reflected in the configuration information. - The operation-unit configuration switching method according to claim 4, wherein
using said switching condition, an operation process time corresponding to operations to be performed by the operation units (10) is calculated according to information about the operations to be performed in each of the plurality of operation units (10) and information about connections between the operation units, an operation unit (10) to be connected first is selected based on the information about connections between the operation units (10), and an accumulated process time is calculated based on the operation process time of each operation unit (10) in connection order starting with the first operation unit (10),
each time the accumulated process time is calculated, the accumulated process time is compared with a predetermined operation cycle, a route is set so as not to use the flip-flop (12) if the accumulated process time is below the operation cycle, a route is set so as to use the flip-flop (12) if the accumulated process time exceeds the operation cycle, and the accumulated time is set to the operation process time of the currently selected operation unit (10). - The operation-unit configuration switching method according to claim 5, wherein the operation process time is calculated in relation to the number of wires connected to the operation unit (10).
- A re-configurable circuit (1) according to claim 3, wherein the operation units are organized into groups, and further comprising
a communication unit provided with a flip-flop (12) for communication at an output of a said group when transferring data between groups if a sum of operation process times of the operation units (10) is below the predetermined operation cycle when connecting the operation units (10) provided with a switching unit for performing the switching, based on the configuration information, for transferring data via the flip-flop (12) for communication. - The operation-unit configuration switching method according to claim 5, wherein
the operation process time further includes time caused by temperature fluctuations.
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2005067294A JP4444860B2 (en) | 2005-03-10 | 2005-03-10 | Reconfigurable circuit and configuration method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
EP1701251A1 true EP1701251A1 (en) | 2006-09-13 |
EP1701251B1 EP1701251B1 (en) | 2009-11-18 |
Family
ID=36190609
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP05256002A Expired - Fee Related EP1701251B1 (en) | 2005-03-10 | 2005-09-27 | Reconfigurable circuit and configuration switching method |
Country Status (4)
Country | Link |
---|---|
US (1) | US7315933B2 (en) |
EP (1) | EP1701251B1 (en) |
JP (1) | JP4444860B2 (en) |
DE (1) | DE602005017735D1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014078753A1 (en) * | 2012-11-19 | 2014-05-22 | Qualcomm Technologies, Inc. | Automatic pipeline stage insertion |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4861030B2 (en) * | 2006-03-24 | 2012-01-25 | 株式会社東芝 | Semiconductor device |
JP6786955B2 (en) * | 2016-08-25 | 2020-11-18 | 富士ゼロックス株式会社 | Reconfigurable logic circuit |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0798633A2 (en) * | 1996-03-29 | 1997-10-01 | Matsushita Electric Industrial Co., Ltd. | Data processor having a variable number of pipeline stages |
US6292926B1 (en) * | 1997-07-03 | 2001-09-18 | Matsushita Electric Industrial Co., Ltd. | Functional module model, pipelined circuit synthesis and pipelined circuit device |
JP2002076883A (en) | 2000-06-15 | 2002-03-15 | Nec Corp | Re-constructable device having programmable interconnecting network suitable for data path |
US20040267994A1 (en) * | 2003-06-12 | 2004-12-30 | Arm Limited | Flexibility of design of a bus interconnect block for a data processing apparatus |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5869979A (en) * | 1996-04-05 | 1999-02-09 | Altera Corporation | Technique for preconditioning I/Os during reconfiguration |
JP2003058426A (en) * | 2001-08-21 | 2003-02-28 | Sony Corp | Integrated circuit, and its circuit constituting method and program |
JP4104538B2 (en) * | 2003-12-22 | 2008-06-18 | 三洋電機株式会社 | Reconfigurable circuit, processing device provided with reconfigurable circuit, function determination method of logic circuit in reconfigurable circuit, circuit generation method, and circuit |
JP4120631B2 (en) * | 2004-10-05 | 2008-07-16 | 株式会社日立製作所 | Semiconductor integrated circuit |
DE102004056738B3 (en) * | 2004-11-24 | 2006-05-18 | Infineon Technologies Ag | Programmable logic cell for a programmable logic array, arithmetic unit and digital circuitry |
JP2006172219A (en) * | 2004-12-16 | 2006-06-29 | Toshiba Corp | Automated design device for semiconductor integrated circuit, automated design method for semiconductor integrated circuit, and automated design program for semiconductor integrated circuit |
JP4527571B2 (en) * | 2005-03-14 | 2010-08-18 | 富士通株式会社 | Reconfigurable processing unit |
JP4675714B2 (en) * | 2005-08-17 | 2011-04-27 | 富士通セミコンダクター株式会社 | Reconfigurable integrated circuit device that automatically builds an initialization circuit |
-
2005
- 2005-03-10 JP JP2005067294A patent/JP4444860B2/en not_active Expired - Fee Related
- 2005-09-27 DE DE602005017735T patent/DE602005017735D1/en active Active
- 2005-09-27 EP EP05256002A patent/EP1701251B1/en not_active Expired - Fee Related
- 2005-10-06 US US11/244,043 patent/US7315933B2/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0798633A2 (en) * | 1996-03-29 | 1997-10-01 | Matsushita Electric Industrial Co., Ltd. | Data processor having a variable number of pipeline stages |
US6292926B1 (en) * | 1997-07-03 | 2001-09-18 | Matsushita Electric Industrial Co., Ltd. | Functional module model, pipelined circuit synthesis and pipelined circuit device |
JP2002076883A (en) | 2000-06-15 | 2002-03-15 | Nec Corp | Re-constructable device having programmable interconnecting network suitable for data path |
US20040267994A1 (en) * | 2003-06-12 | 2004-12-30 | Arm Limited | Flexibility of design of a bus interconnect block for a data processing apparatus |
Non-Patent Citations (6)
Title |
---|
AMANO H ET AL: "A DYNAMICALLY ADAPTIVE HARDWARE ON DYNAMICALLY RECONFIGURABLE PROCESSOR", IEICE TRANSACTIONS ON COMMUNICATIONS, COMMUNICATIONS SOCIETY, TOKYO, JP, vol. E86-B, no. 12, December 2003 (2003-12-01), pages 3385 - 3391, XP001191542, ISSN: 0916-8516 * |
BROWN S ET AL: "FPGA AND CPLD ARCHITECTURES: A TUTORIAL", IEEE DESIGN & TEST OF COMPUTERS, IEEE SERVICE CENTER, NEW YORK, NY, US, vol. 13, no. 2, 1 June 1996 (1996-06-01), pages 42 - 57, XP000596695, ISSN: 0740-7475 * |
EBELING C ET AL: "Mapping applications to the RaPiD configurable architecture", ANNUAL IEEE SYMPOSIUM ON FIELD-PROGRAMMABLE CUSTOM COMPUTING MACHINES, 1997, pages 106 - 115, XP002132822 * |
HENNESSYJ ET AL: "Computer Architecture - A quantative approach", COMPUTER ARCHITECTURE A QUANTITATIVE APPROACH, SAN MATEO, MORGAN KAUFMANN, US, 1990, pages 250 - 273, XP002297069 * |
MUNZER A ET AL: "Converting combinational circuits into pipelined data paths", INTERNATIONAL CONFERENCE ON COMPUTER AIDED DESIGN. (ICCAD). SANTA CLARA, NOV. 11 - 14, 1991, LOS ALAMITOS, IEEE. COMP. SOC. PRESS, US, vol. CONF. 9, 11 November 1991 (1991-11-11), pages 368 - 371, XP010023677, ISBN: 0-8186-2157-5 * |
NOTE S ET AL: "COMBINED HARDWARE SELECTION AND PIPELINING IN HIGH-PERFORMANCE DATA-PATH DESIGN", IEEE TRANSACTIONS ON COMPUTER AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 11, no. 4, 1 April 1992 (1992-04-01), pages 413 - 423, XP000268824, ISSN: 0278-0070 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014078753A1 (en) * | 2012-11-19 | 2014-05-22 | Qualcomm Technologies, Inc. | Automatic pipeline stage insertion |
US9110689B2 (en) | 2012-11-19 | 2015-08-18 | Qualcomm Technologies, Inc. | Automatic pipeline stage insertion |
Also Published As
Publication number | Publication date |
---|---|
US7315933B2 (en) | 2008-01-01 |
JP2006252144A (en) | 2006-09-21 |
US20060202715A1 (en) | 2006-09-14 |
EP1701251B1 (en) | 2009-11-18 |
DE602005017735D1 (en) | 2009-12-31 |
JP4444860B2 (en) | 2010-03-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8356063B2 (en) | Reconfigurable digital filter | |
US5778251A (en) | State machine in which subsequent state can be directly inputted externally and transition can be stopped retaining the current state | |
US7167727B2 (en) | Wireless data communications using FIFO for synchronization memory | |
US8138788B2 (en) | Reconfigurable device | |
US20060036974A1 (en) | IP-based LSI design system and design method | |
US20090282213A1 (en) | Semiconductor integrated circuit | |
JP2002251232A (en) | Logical circuit module having power consumption control interface and storage medium with the same module stored | |
KR100542118B1 (en) | A digital intermediate-frequency signal processor and digital filter supporting software-defined radio system and its design | |
US6604232B2 (en) | High-level synthesis method and storage medium storing the same | |
JP3351452B2 (en) | Programmable gate array | |
EP1701251A1 (en) | Reconfigurable circuit and configuration switching method | |
US5835734A (en) | Electronic processing and control system with programmable hardware | |
EP0843893A1 (en) | A microcontroller having an n-bit data bus width with less than n i/o pins and a method therefor | |
JP2006236106A (en) | Data processor and data processing method | |
US6938232B2 (en) | Floorplanning apparatus deciding floor plan using logic seeds associated with hierarchical blocks | |
US7742469B2 (en) | Data input circuit and semiconductor device utilizing data input circuit | |
EP4120073A1 (en) | Configuration method for logical address space of hardware single board, device, and storage medium | |
US7143302B2 (en) | Pipeline structure | |
JP3499460B2 (en) | Spreading code generating circuit and spreading code generating method | |
US20100284234A1 (en) | Memory control method and memory control device | |
US7996657B2 (en) | Reconfigurable computing circuit | |
EP2015525A2 (en) | Data transmitting apparatus, data transmitting method and recording medium including data transmission program | |
CN113407388B (en) | FPGA (field programmable Gate array) online debugging method with dynamically variable sampling frequency | |
CN117422024B (en) | Data bit width conversion method, device, computer equipment and medium | |
JP5153642B2 (en) | Dynamic constant folding of circuits |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC NL PL PT RO SE SI SK TR |
|
AX | Request for extension of the european patent |
Extension state: AL BA HR MK YU |
|
17P | Request for examination filed |
Effective date: 20070307 |
|
17Q | First examination report despatched |
Effective date: 20070410 |
|
AKX | Designation fees paid |
Designated state(s): DE FR GB |
|
GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
GRAS | Grant fee paid |
Free format text: ORIGINAL CODE: EPIDOSNIGR3 |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Kind code of ref document: B1 Designated state(s): DE FR GB |
|
REG | Reference to a national code |
Ref country code: GB Ref legal event code: FG4D |
|
REF | Corresponds to: |
Ref document number: 602005017735 Country of ref document: DE Date of ref document: 20091231 Kind code of ref document: P |
|
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
26N | No opposition filed |
Effective date: 20100819 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: DE Payment date: 20140923 Year of fee payment: 10 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: GB Payment date: 20140924 Year of fee payment: 10 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: FR Payment date: 20140906 Year of fee payment: 10 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R119 Ref document number: 602005017735 Country of ref document: DE |
|
GBPC | Gb: european patent ceased through non-payment of renewal fee |
Effective date: 20150927 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: ST Effective date: 20160531 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20160401 Ref country code: GB Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20150927 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: FR Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20150930 |